A comparison of the usability of performing ad hoc querying on dimensionally modeled data versus operationally modeled data

To study the usability of dimensionally modeled data (DMD), we conducted an experiment involving 113 participants to analyze their performance in ad hoc querying tasks. The results show that DMD's usability is high according to both objective and subjective measurements. Usability is dependent upon the segment of data model used and, in particular, on the type of task being performed, with the highest results for aggregations and holistic view creation. Interestingly, dimensionally modeled data seem to be distinguished by learnability. The study's results partially agree with the results obtained in usability studies of relational data models and similar data structures.

► Dimensionally modeled data were compared with operationally modeled data.
► Business user's performances in ad hoc querying were measured.
► Dimensionally modeled data improve accuracy and shorten completion time.
► Most of separately analyzed dimensional modeling innovations improve usability.
► A better holistic view is gained from dimensionally modeled data.